7The technological challenges facing agriculture

Transcription

1 Innovating through science and technology chapter 158 7The technological challenges facing agriculture in the 21st century are probably even more daunting than those in recent decades. With the increasing scarcity of land and water, productivity gains will be the main source of growth in agriculture and the primary means to satisfy increased demand for food and agricultural products. With globalization and new supply chains, farmers and countries need to continually innovate to respond to changing market demands and stay competitive. With climate change, they will have to gradually adapt. All regions, especially the heterogeneous and risky rainfed systems of Sub-Saharan Africa, need sustainable technologies that increase the productivity, stability, and resilience of production systems. 1 These changes imply that technology for development must go well beyond just raising yields to saving water and energy, reducing risk, improving product quality, protecting the environment, and tailoring to gender differences. Science is also changing rapidly. Revolutionary advances in the biological and information sciences have the potential to enhance the competitiveness of marketoriented smallholders and overcome drought and disease in production systems important to the poor. Consider the winwin-win of transgenic insect-resistant cotton: it has reduced yield losses, increased farmer profits, and greatly reduced pesticide use for millions of smallholders. But the benefits of biotechnology, driven by large, private multinationals interested in commercial agriculture, have yet to be safely harnessed for the needs of the poor. The institutional setting for technological innovation is changing rapidly as well it is more complex, involving plural systems and multiple sources of innovation. The new world of agriculture is opening space for a wider range of actors in innovation, including farmers, the private sector, and civil society organizations. Linking technological progress with institutional innovations and markets to engage this diverse set of actors is at the heart of future productivity growth. These changes focus attention on wider innovation systems. With the development of markets, innovation becomes less driven by science (supply side) and more by markets (demand side). New demand-driven approaches stress the power of users men and women farmers, consumers, and interests outside of agriculture in setting the research agenda and the importance of research in a value chain from farm to plate. Innovation for the new agriculture requires feedback, learning, and collective action among this much broader set of actors. This chapter looks at the recent record of science and technological innovation from three perspectives: The recent impacts and emerging challenges of biological and management technologies The investments in research and development (R&D) to generate new technologies, paying particular attention to growing divides between industrial and developing countries, and within the developing countries themselves The emerging institutional arrangements that make investments in innovation, including extension, more efficient and effective in meeting market demands through collective action and farmer involvement The main conclusion: Investments in agricultural R&D have turned much

2 Innovating through science and technology 159 of developing-world agriculture into a dynamic sector, with rapid technological innovation accelerating growth and reducing poverty. But global and national market failures continue to induce serious underinvestment in R&D and in related extension systems, especially in the agriculture-based countries of Africa. Increasing public and private investment in R&D and strengthening institutions and partnerships with the private sector, farmers, and civil society organizations are now essential to assess user demand for R&D, increase market responsiveness and competitiveness, and ensure that the poor benefit. These investments and institutional innovations will be even more important in the future, with rapidly changing markets, growing resource scarcity, and greater uncertainty. Genetic improvement has been enormously successful, but not everywhere Agriculture is a biological process so technological innovation in agriculture is different from that in other sectors. The 1950s and 1960s showed that genetic improvement technologies such as crop and animal breeds were often location specific and generally did not travel well from the temperate North to the tropical South. Research since the 1960s aimed at adapting improved varieties and animal breeds to subtropical and tropical conditions has generated high payoffs and pro-poor impacts. Rapid advances in the biological and informational sciences promise even greater impacts that have yet to be tapped for the benefit of the poor (see focus E). Slow magic: the continuing spread of improved varieties Since the 1960s, scientific plant breeding that developed improved varieties suited to smallholders in subtropical and tropical areas the green revolution has been one of the major success stories of development (figure 7.1). Initially spearheaded by semidwarf varieties of rice and wheat and improved varieties of maize from international agricultural research centers of the Consultative Group on International Agricultural Research (CGIAR), public breeding programs in developing countries have released more than 8,000 improved crop varieties over the past 40 years. 2 Private seed companies have also become significant sources of improved hybrid varieties for smallholders for some crops, especially maize. The contribution of improved crop varieties to yield growth since 1980 has been even greater than in the green revolution Figure 7.1 Improved varieties have been widely adopted, except in Sub-Saharan Africa Area planted with improved varieties, , % of crop area Rice Wheat Maize Sorghum Cassava Potatoes Sub-Saharan Africa South Asia East Asia & Pacific Middle East & North Africa Latin America & Caribbean Sources: WDR 2008 team, based on Evenson 2003; IRRI, personal communication 2007; CIMMYT, personal communication Notes: Improved varieties of rice and wheat are semidwarf varieties first developed in what became known as the green revolution. Data are provided for the period , except for maize in some Sub-Saharan African countries where data are from 1997.

3 160 WORLD DEVELOPMENT REPORT 2008 decades. In the 1980s and 1990s, improved varieties are estimated to have accounted for as much as 50 percent of yield growth, compared with 21 percent in the preceding two decades. Poor consumers have been the main beneficiaries. Without those gains in yields, world cereal prices would have been percent higher in 2000, caloric availability per capita in developing countries would have been 4 7 percent lower, million more children would have been classified as malnourished, and many more hectares of forest and other fragile ecosystems would have been brought under cultivation. 3 Steady genetic improvements to newer generations of varieties and their spread beyond irrigated areas and rainfed areas with good water control have contributed to continuing yield gains. For example, improved varieties are now planted on 80 percent of the cereal area in India, only about half of it irrigated. 4 Newer generations of improved wheat varieties have provided an annual increase in yields of 1 percent, and globally the area planted with them has more than doubled since 1981, largely in rainfed areas. 5 Not all farmers have been touched by this slow magic. 6 Sub-Saharan Africa has seen very incomplete adoption, with many countries having almost no area under improved varieties. Why the limited green revolution in Sub-Saharan Africa? 7 The broader mix of crops grown in the region; the agroecological complexities and heterogeneity of the region; the lack of infrastructure, markets, and supporting institutions; and the gender differences in labor responsibility and access to assets all have contributed (chapter 2). 8 Recent experience in Sub-Saharan Africa offers more promise. After a late start, improved varieties are finally making an impact on some food staples: Maize. Improved maize varieties and hybrids were widely adopted by smallholders in many African countries in the 1980s, reaching almost universal coverage in a few countries, such as Zimbabwe. But much of this was underwritten by heavy subsidies for inputs and prices, subsidies that were unsustainable. 9 Still, a substantial share of the maize area was planted to improved varieties and hybrids in 2006 in Kenya (80 percent), Malawi (30 percent), Tanzania (28 percent), Zambia (49 percent), and Zimbabwe (73 percent). 10 Cassava. Improved disease-resistant strains of cassava have been adopted, reaching more than half the cassava area in Nigeria, the world s largest producer. Cassava has been the fastest growing food staple in Africa, and since it is a staple of the poor, the impacts of productivity gains are especially pro-poor. 11 Rice. The New Rice for Africa combining the high-yielding potential of Asian rice with the resistance of African rice to weeds, pests, diseases, and water stress was released to farmers in Increasing yields under low input conditions, it is cultivated on about 200,000 hectares in Africa. 12 Yet adoption is still modest because of insufficient dissemination, training, and extension. Beans. In eastern, central, and southern Africa, nearly 10 million farmers, mostly women, are reportedly growing and consuming new bean varieties (Phaseolus vulgaris), many with multiple stress resistances. 13 A complementary institutional development in low and uncertain rainfall regions of marginal production potential is participatory varietal selection and breeding approaches that involve farmers in the early stages of plant breeding. Decentralized and participatory approaches allow farmers to select and adapt technologies to local soil and rainfall patterns and to social and economic conditions, using indigenous knowledge as well. Between 1997 and 2004, the Barley Research Program of the International Center for Agricultural Research in Dry Areas in Syria transformed its operation from 8,000 plots planted and evaluated on the research station to 8,000 plots planted in farmers fields and evaluated by farmers. 14 It was found that participatory plant breeding and varietal selection speeds varietal development and dissemination to 5 7 years, half the years in a conventional plant-breeding program. 15

4 Innovating through science and technology 161 In the very poor, rainfed rice-growing areas of South Asia that the green revolution passed by, participatory plant breeding is now paying off with strong early adoption of farmer-selected varieties that provide 40 percent higher yields in farmers fields. 16 The approach needs to be more widely tested in the heterogeneous rainfed environments of Africa, where involving farmers, especially women farmers, in selecting varieties has shown early successes for beans, maize, and rice. 17 The cost effectiveness of the approach for wider use also needs to be evaluated. But improved varieties alone will not produce a green revolution in less-favored areas; low soil fertility and lack of water control are major constraints that are difficult to overcome through genetic enhancement alone. In the language of crop scientists, both the G (genotype) and the E (crop environment and management) have to change to exploit the type of positive G E interactions that characterize a green revolution. Yield risk and the Red Queen Yield stability is important for all farmers, but especially for subsistence-oriented farmers whose food security and livelihood are vulnerable to pest and disease outbreaks, droughts, and other stresses. Improved varieties can make yields more stable. A recent study concluded that the variability of cereal yields, measured by the coefficient of variation around trends over the past 40 years, has declined in developing countries, a decline that is statistically associated with the spread of improved varieties, even after controlling for more irrigation and other inputs. 18 The annual benefits from better yield stability in maize and wheat alone are estimated at about $300 million more than the annual spending on maize- and wheat-breeding research in the developing world. Yield stability of improved varieties largely reflects long-standing efforts in breeding for disease and pest resistance. Even when improved varieties are bred to resist a disease, they must be periodically replaced to ensure against outbreaks from new races of pathogens. Without investment in such maintenance research, yields would decline a situation best described by the Red Queen in Alice in Wonderland: Now here, you see, it takes all the running you can do to keep in the same place. 19 A third to a half of current R&D investments in crop breeding may be for maintenance, leaving reduced resources to address productivity advances. 20 Underinvesting in maintenance research can threaten local food supplies and sometimes have global significance. Consider the dramatic recent emergence of Ug99, a new race of stem rust (Puccinia graminis tritici) in wheat, the world s second most important food staple. Stem rust is catastrophic because it can cause an almost complete loss of crops over wide areas. Ug99 first appeared in 1999 in Uganda and is now widespread in wheat-growing areas of Kenya and Ethiopia; in 2007 it was found in Yemen. Based on previous experience, Ug99 is expected to be carried by the wind through the Middle East to wheat-growing areas of South Asia and possibly to Europe and the Americas. Given the narrow base of genetic resistance to the disease in existing varieties of wheat, the spread of Ug99 could cause devastating losses in some of the world s breadbaskets. 21 The last major outbreak of stem rust in the United States in 1953 and 1954 caused a 40 percent yield loss worth $3 billion in today s dollars. 22 Through a new international effort, plant breeders and pathologists should be able to avoid a global epidemic by screening for resistant genotypes and getting them into farmers fields. Farmers who use traditional varieties are also vulnerable to random outbreaks of disease, as with the recent outbreak of bacterial wilt (Banana Xanthomonas wilt) in East Africa. The disease threatens the livelihoods and food security of millions of people who depend on bananas in the Great Lakes Region an area that boasts the world s highest per capita consumption of bananas. 23 In Uganda, where bananas are a staple, the potential national loss is estimated at $360 million a year. 24 A genetically engineered variety with resistance to the disease is a breakthrough, but applying it depends on Uganda s putting biosafety regulations in place (see focus E). 25

5 162 WORLD DEVELOPMENT REPORT 2008 These recurring crises are wake-up calls to develop appropriate maintenance research strategies together with global coordination, surveillance, and financing. Progress in developing varieties that perform well under drought, heat, flood and salinity has been generally slower than for disease and pest resistance. The International Maize and Wheat Improvement Center (CIMMYT), after more than 30 years of research to produce drought-tolerant maize varieties and hybrids, is now seeing results in eastern and southern Africa. Evaluated against existing hybrids, the new ones yield 20 percent more on average under drought conditions. 26 Similarly, recent evidence points to significant yield gains in breeding wheat for drought and heat-stressed environments. 27 New varieties of rice that survive flooding have also been identified. 28 Such advances in drought, heat, and flood tolerance will be especially important in adapting to climate change. But large areas of major food crops are now planted each year in relatively few improved varieties, and genetic uniformity can make crops vulnerable to major yield losses. There is some evidence that genetic uniformity increases yield risk, even though it can also produce higher yields. 29 In recent decades, the world has largely avoided major disasters from genetic uniformity, in part because of frequent turnover of varieties, which brings new sources of resistance. Even so, wider conservation and use of genetic resources are needed (chapter 11). Beyond crops: genetic improvement of livestock and fish Advances in animal and fish genetics combined with improved animal health and feeding have been the basis of the livestock revolution in developing countries (chapter 2). Improved pig and poultry breeds have been adopted through private direct transfers from the North. 30 These gains show up in livestock productivity. Over in the developing world, the annual off-take from a flock of chickens with a total live weight of 1,000 kilograms increased from 1,290 kilograms to 1,990 kilograms and that of pigs improved from 140 kilograms to 330 kilograms live weight. 31 The cross-breeding of dairy cows with exotic breeds has improved the livelihoods of smallholder farmers in high-potential areas in the tropics. About 100 million cattle and pigs are bred annually in the developing world using artificial insemination. 32 And thanks largely to artificial insemination, about 1.8 million small-scale farmers in the highlands of East Africa draw a significant part of their livelihood from the higher milk yields they obtain from genetically improved dairy cattle. 33 Similarly for fish, genetically improved tilapia is changing aquaculture into one of the fastest growing sectors in Asian agriculture. In 2003 improved strains from a single project for the genetic improvement of farmed tilapia (GIFT) accounted for 68 percent of the total tilapia seed produced in the Philippines, 46 percent in Thailand, and 17 percent in Vietnam. Lower production costs per kilogram of fish, high survival rates, higher average weight per fish, and yields 9 54 percent higher than existing strains explain the fast uptake of GIFTderived strains. 34 Even so, genetic improvement in animals and fish have reached only a small share of developing-country farmers, partly because of constraints in the delivery systems for these technologies. Livestock breeding services in much of the developing world are still generally subsidized, crowding out the private sector. More research to reduce the costs of these technologies, and more policy and institutional reforms to ensure more efficient and widespread delivery, will enable the developing world to capture the full benefits of these promising technologies. A biotechnology revolution in the making? Agricultural biotechnology has the potential for huge impacts on many facets of agriculture crop and animal productivity, yield stability, environmental sustainability, and consumer traits important to the poor. The first-generation biotechnologies include plant tissue culture for micropropagation and production of virus-free planting materials, molecular diagnostics of crop and livestock diseases, and embryo transfer in livestock. Fairly cheap and eas-

6 Innovating through science and technology 163 ily applied, these technologies have already been adopted in many developing countries. For instance, disease-free sweet potatoes based on tissue culture have been adopted on 500,000 hectares in Shandong Province in China, with yield increases of percent, 35 and advanced biotechnology-based diagnostic tests helped eradicate rinderpest virus in cattle. The second-generation biotechnologies based on molecular biology use genomics to provide information on genes important for a particular trait. This allows the development of molecular markers to help select improved lines in conventional breeding (called marker-assisted selection). Such markers are speeding the breeding, leading to downy mildew resistant millet in India; cattle with tolerance to African sleeping sickness; and bacterial leaf blightresistant rice in the Philippines. 36 As the costs of marker-assisted selection continues to fall, it is likely to become a standard part of the plant breeder s toolkit, substantially improving the efficiency of conventional breeding. The most controversial of the improved biotechnologies are the transgenics, or genetically modified organisms, commonly known as GMOs (see focus E). Transgenic technology is a tool for precision breeding, transferring a gene or set of genes conveying specific traits within or across species. About 9 million smallholder farmers, mainly in China and India, have adopted transgenic Bt cotton for insect resistance. It has already reduced yield losses from insects, increased farmer s profits, and significantly reduced pesticide use in India and China. Transgenic technology remains controversial, however, because of perceived and potential environmental and health risks. Biotechnology thus has great promise, but current investments are concentrated largely in the private sector, driven by commercial interests, and not focused on the needs of the poor. That is why it is urgent to increase public investments in propoor traits and crops at international and national levels and to improve the capacity to evaluate the risks and regulate these technologies in ways that are cost effective and inspire public confidence in them. The potential benefits of these technologies for the poor will be missed unless the international development community sharply increases its support to interested countries (see focus E). Management and systems technologies need to complement genetic improvement Much R&D is focused on improving the management of crop, livestock, and natural resource systems. The CGIAR invests about 35 percent of its resources in sustainable production systems, twice the 18 percent it invests in genetic improvement. 37 Much of this work has emphasized soil and water management and agroecological approaches that exploit biological and ecological processes to reduce the use of nonrenewable inputs, especially agricultural chemicals. 38 Examples include conservation tillage, improved fallows and soils, green manure cover crops, soil conservation, and pest control using biodiversity and biological control more than pesticides. Zero tillage One of the most dramatic technological revolutions in crop management is conservation (or zero) tillage, which minimizes or eliminates tillage and maintains crop residues as ground cover. It has many advantages over conventional tillage: increasing profitability from savings in labor and energy, conserving soil, increasing tolerance to drought, and reducing greenhouse gas emissions. But it makes the control of weeds, pests, and diseases more complex, and it usually requires some use of herbicides. In Latin America (mainly Argentina and Brazil), zero tillage is used on more than 40 million hectares (about 43 percent of the arable land). 39 Originally adopted by large and midsize farmers, the practice has spread to small farmers in southern Brazil. Networks of researchers, input suppliers, chemical companies, and farmers have used participatory research and formal and informal interactions to integrate various parts of the technology (rotations, seeds,

7 164 WORLD DEVELOPMENT REPORT 2008 BOX 7.1 When zero means plenty: the benefits of zero tillage in South Asia s rice-wheat systems South Asia s rice-wheat systems, the bedrocks of food security, are in trouble (chapter 8). Long-term experiments show that crop yields are stagnating and that soil and water quality are in decline. In response, the Rice Wheat Consortium of the Indo- Gangetic Plain of South Asia a network of international scientists, national scientists, extension agents, private machinery manufacturers, and nongovernmental organizations (NGOs) has developed and promoted zero-tillage farming. Although zero tillage is part of a much broader farm management system that involves many agricultural practices, a key part of the system promoted by the consortium is planting wheat immediately after rice without tillage so that the wheat seedlings germinate using the residual moisture from the previous rice crop. A notable aspect of the approach has been to work with local machinery manufacturers and farmers to adapt drills to local conditions. Zero-tillage farming increases wheat yields through timely sowing and reduces production costs by up to 10 percent. It reduces water use by about 1 million liters per hectare (a saving of percent). It improves soil structure, fertility, and biological properties and reduces the incidence of weeds and some other pests. Zero tillage with wheat succeeding rice is now the most widely adopted resource-conserving technology in the Indo-Gangetic Plain, especially in India with some 0.8 million hectares planted in 2004 using the method. Research on zero tillage on rice-wheat systems in India is estimated to have a rate of return of 57 percent, based on an investment of $3.5 million. 40 Further work must consider the fact that women contribute more than half the labor in the rice-wheat system, especially for livestock management. This has important implications for involving women in seed selection and fodder management practices for the system. Sources: Malik, Yadav, and Singh 2005; Paris chemicals, and machinery) and adapt them to local conditions. The approach was also used by an estimated 100,000 smallholders in Ghana in the past decade. 41 It is also being rapidly adopted in the irrigated wheat-rice systems of the Indo-Gangetic Plain (box 7.1). Legumes and soil fertility Another input-saving and resourceconserving technology is introducing or improving legumes in farming systems to provide multiple benefits, most notably biologically fixing nitrogen that reduces the need for chemical fertilizer (especially if the legume is inoculated with nitrogenfixing Rhizobium). Much of the yield gain in Australian cereal production over the past 60 years comes from rotation systems that include legumes. 42 In southern Africa, fast-growing fertilizer trees such as Gliricidia, Sesbania, and Tephrosia have improved soil fertility, soil organic matter, water infiltration, and holding capacity. Other benefits include reduced soil erosion and the production of fuelwood and livestock fodder (box 7.2). 43 These technologies are quite location specific, however, and research to adapt them to farming systems defined by soils, land pressure, and labor availability (differentiated by men and women) should be a high priority to address the severe depletion of soil nutrients in Sub-Saharan Africa. Pest management At the other end of the spectrum, research that reduces use of dangerous pesticides can have win-win-win benefits for profitability, the environment, and human health in intensive systems. Integrated pest management uses a combination of practices, especially improved information on pest populations and predators to estimate pest losses and adjust pesticide doses accordingly. Despite notable examples of integrated pest management, adoption has often been limited because of its complexity (chapter 8). However, biological control of pests can sometimes have spectacular impacts, often requiring no action on the part of farmers. One of the best-documented cases is the control of the cassava mealybug in Sub- Saharan Africa, which was introduced accidentally with planting material from Latin America in the 1970s, causing significant economic losses. 44 The International Institute for Tropical Agriculture responded to the crisis by selecting, rearing, and distributing in 20 countries a parasitoid wasp that was the mealybug s natural enemy. The biological control provided by the wasp was so effective that the cassava mealybug is now largely controlled. Even when using the most conservative assumptions, the return on this research investment has been extremely high (net present value estimated at US$9 billion). 45 Combinations The greatest impact on productivity is obtained through production ecology approaches that combine improved varieties and several management technologies, crop-livestock integration, and mechanical technologies to exploit their synergistic effects. 46 For example, in Ghana zero tillage is combined with improved legume-based

8 Innovating through science and technology 165 fallows and maize varieties. 47 In eastern Africa, low-input integrated pest management has been developed by planting Desmodium (a nitrogen-fixing leguminous plant that can be used for livestock fodder) between the rows of maize to suppress Striga, an especially serious parasitic weed. 48 A similar integrated approach involving improved varieties, biological nitrogen fixation, cover crops, and machinery adapted to zero tillage has been vital to the global competitiveness of Brazilian soybeans. 49 With the rise of value chains, such technologies must also often integrate product quality and agricultural processing. The need for more suitable technologies Although R&D on production and resource management has huge potential, success has been mixed, with zero tillage as the outstanding success. Suitable technologies are still badly needed to conserve and efficiently use scarce water, control erosion, and restore soil fertility for smallholders in less-favored areas. However, such complex technologies are often labor or land intensive and may be unattractive to farmers where labor costs are high, land is scarce, or discount rates on future returns are very high or the returns risky. These concerns are especially important to women farmers lacking access to assets and services and who have specific seasonal labor-use patterns. Although the technologies are aimed at poor farmers, the record shows higher adoption levels by wealthier farmers. 50 Management and systems technologies can require considerable institutional support to be widely adopted (chapter 8). Many of them involve the interaction of several actors such as collective action among neighboring farmers as well as technical support, learning, farmer-to-farmer interaction, and knowledge sharing, as with conservation tillage in Brazil. In addition, many technologies have positive impacts on the environment that are not captured in the private benefits for adopting farmers and may require payment for environmental services to encourage their adoption (chapter 8). BOX 7.2 Using legumes to improve soil fertility The low fertility in much of African soil and the low (and sometimes declining) use of mineral fertilizers have increased farmer interest in agroforestry-based soil fertility systems. The main methods are a rotational fallow or a permanent intercrop of nitrogen-fixing trees. The systems have spread mainly in the southern African subhumid region, where they have more than doubled maize yields and increased net returns on land and labor. In Zambia, the financial benefits to the nearly 80,000 farmers practicing improved fallows were The integrative nature of management and agroecological approaches also affects the way R&D is carried out. Because of location specificity, farmer and community participation in R&D characterizes the major success stories of these technologies. Location specificity also reduces the potential for spillovers of technologies from other regions so despite substantial investment by the CGIAR, the evidence of impacts is limited. 51 For these reasons, scaling up management and system technologies will not be easy. Networks of scientists, farmers, private firms, and NGOs take time to develop and become inclusive and effective. They also take time to develop the ecological literacy to successfully apply many of these technologies (chapter 8). But advances in geographic information systems and remote sensing by satellites are opening new ways to synthesize complex and diverse spatial data sets, creating new opportunities for collaboration among scientists, policy makers, and farmers. Investing more in R&D Agricultural productivity improvements have been closely linked to investments in agricultural R&D (chapter 2). 52 Published estimates of nearly 700 rates of return on R&D and extension investments in the developing world average 43 percent a year. 53 Returns are high in all regions, including Sub-Saharan Africa (figure 7.2). Even discounting for selection bias in evaluation studies and other methodological almost $2 million for 2005/06. The technologies often work best in combination with judicious doses of mineral fertilizer. With 12 million smallholder maize farmers in eastern and southern Africa, rotational fallows and permanent intercropping offer considerable long-term opportunities for integrated soil fertility management to keep African soils productive and healthy. Source: Consultative Group on International Agricultural Research Science Council (CGIAR) 2006a.

9 166 WORLD DEVELOPMENT REPORT 2008 Figure 7.2 Estimated returns to investment in agricultural R&D are high in all regions a averaging 43 percent All countries (1673) All developed countries (990) All developing countries (683) Sub-Saharan Africa (188) Asia (222) Middle East & North Africa (11) Latin America & Caribbean (262) Percent Source: Alston and others a. Based on studies carried out from 1953 to Number of observations in parentheses. issues, 54 there is little doubt that investing in R&D can be a resounding success. The high payoffs relative to the cost of capital also indicate that agricultural science is grossly underfunded. Why agricultural R&D is underfunded Public investment is especially important for funding agricultural R&D where markets fail because of the difficulty of appropriating the benefits. Seeds of many improved varieties can be reused by farmers and sold or shared with neighboring farmers (nonexcludable). Information on improved management practices can be freely exchanged (nonrival). Intellectual property rights (IPRs) have partially overcome these market failures in industrial countries, but few technologies of importance to poor farmers can be cost-effectively protected by IPRs (box 7.3). A major exception is private sector investment in hybrid seed of a few crops where intellectual property can be protected by trade secrets. Farmers must purchase hybrid seed frequently to maintain its yield advantage, providing a steady market for private seed companies. Star performers and the others. For these reasons, private investment in developing-country R&D has been very limited 94 percent of the agricultural R&D in the developing world is conducted by the public sector. 55 But even growth in public spending on R&D, after rapidly increasing in the 1960s and 1970s, has slowed sharply in most regions in the past decade or more, opening a knowledge divide between poor countries and rich countries and within the developing world between a handful of star performers and most of the others. Developing countries as a group invested 0.56 percent of their agricultural gross domestic product (GDP) in agricultural R&D in 2000 (including donor contributions), only about one-ninth of the 5.16 percent that developed countries invest. Part of this disparity is because private investment makes up just over half of R&D spending in industrial countries but only 6 percent in the developing world. Still, the intensity of public investment (in relation to agricultural GDP) is five times higher in industrial countries (table 7.1). A few developing countries notably China, India, and to a less extent, Brazil have rapidly increased their spending on agricultural R&D over the past two decades. Their shares in developing-country public spending in agricultural R&D increased from a third in 1981 to almost half in Including spending on science and technology for all sectors, these three countries accounted for 63 percent of the total which is meaningful, because an increasing share of agricultural R&D is carried out in general science and technology organizations. 56 The private sector also has a growing presence in these countries, where expanding agricultural input markets provide incentives to invest. Meanwhile, many agriculture-based countries are flagging or slipping in the amount spent on R&D. In the 1990s, public R&D spending in Sub-Saharan Africa fell in nearly half the 27 countries with data, and the share of agricultural GDP invested in R&D fell on average for the whole region. 57 Politics, prices, and spillovers. Why does this underinvestment in R&D continue, given the well-documented high rate of return on investment? Three main reasons: First, the political economy of public expenditure decisions tends to emphasize short-term payoffs and subsidies that are

10 Innovating through science and technology 167 BOX 7.3 Stronger IPRs in developing countries: effect on small farmers Under the World Trade Organization (WTO) Agreement on Trade-Related Aspects of Intellectual Property Rights, member countries are required to implement IPRs, including those for plant varieties and biotechnology inventions. The most common type of protection is through plant variety rights. A handful of developing countries also provide patent protection. Many developing countries have elected to follow the model developed in 1978 by industrial countries, the Convention on the Protection of New Varieties of Plants known by its implementing agency, the International Union for the Protection of New Varieties of Plants (UPOV), which harmonized conditions and norms for protecting new varieties while giving farmers the right to save and exchange seed. Other countries (for example, India and Thailand) explicitly recognize framework farmers rights to save and exchange seed (derived from the 2004 international treaty of the Food and Agriculture Organization of the UN [FAO]) and to share benefits arising from the use of farmers genetic resources and indigenous knowledge (based on the 1993 Convention on Biological Diversity). North-South bilateral and regional trade agreements often put pressure on developing countries to adopt even stronger protection such as that based on the 1991 Convention of UPOV, which makes selling and exchanging seed of protected varieties illegal. Little impact so far A recent review of the impacts of stronger IPRs on the seed industries of China, Colombia, India, Kenya, and Uganda found relatively little impact to date, mainly because the IPRs are still under development in most countries. Although limitations on the exchange of farmer-saved seed appear a significant obstacle to smallholder farmers, there are no indications that such rules have been enforced. Indeed, it is generally not cost effective to enforce such rules for staple crops grown by smallholders. Also, the potential advantages of IPRs should not be overrated in most developing countries. Relative to broader investment climate issues, IPRs do not seem critical in the initial development of a private seed sector, but they could help to support a maturing commercial seed industry. How countries could do more Even so, countries could do more to adapt IPR legislation to their needs within the guidelines of current international treaties. For example, a country could provide strong protection for commercial crops as an incentive for private investment, while excluding or providing weaker protection to staple food crops important to subsistence-oriented farmers, where seed saving and exchange are integral to farming practices. Only a few developing countries with large commercial sectors or potential in private biotechnology R&D should consider strong IPRs, such as UPOV 1991 and strong patent laws. Plant variety rights also need to fit into other regulatory systems, such as seed certification laws, biosafety laws, and such other IPRs as trademarks and trade secrets. In any event, sharply increased capacity of the public sector, private firms, and farmers is needed to design and build credible and cost-effective IPR systems that fit a country s needs. Sources: Oxfam International 2007b; Tripp, Louwaars, and Eaton 2007; World Bank 2006k. Table 7.1 Total public agricultural R&D expenditures by region, 1981 and 2000 Public agricultural R & D spending R & D spending as a % of agricultural GDP int l $, millions Sub-Saharan Africa 1,196 1, Asia & Pacific 3,047 7, China 1,049 3, India 533 1, West Asia & North Africa 764 1, Latin America & Caribbean 1,897 2, Brazil 690 1, Developing countries 6,904 12, Japan 1,832 1, United States 2,533 3, Developed countries 8,293 10, Total 15,197 23, Sources: Agricultural Science and Technology Indicators database, Pardey and others Note: These estimates exclude Eastern Europe and the former Soviet Union countries because data are not available. politically visible (chapter 4), while agricultural R&D investments are both long term (10 years or more) and risky. Moreover, in agriculture-based countries, the political power of farmers is low anyhow (chapter 1). Second, trade distortions and national policies that reduce incentives to farmers in developing countries are a disincentive to both public and private investment in R&D (chapter 4). 58 Third, because the benefits of much public R&D spill over to other countries, it might not make much economic sense for small countries to spend their scarce

11 168 WORLD DEVELOPMENT REPORT 2008 resources on agricultural science, on their own behalf; many nations have been freeriding on the efforts of a few others. The international agricultural research centers of the CGIAR were created specifically to provide spillovers in many areas of technology. 59 Over half of all benefits of R&D are generated by such spillovers. 60 But future reliance on spillovers for productivity enhancement carries risks. 61 Privatization of R&D restricts access to proprietary technologies and the sharing of scientific knowledge (see below). Traditional sources of spillovers for productivity growth the public R&D systems in developed countries and the CGIAR have also shifted priorities away from productivity-enhancing research to research on the environment and food safety and quality. 62 In some regions, especially Sub- Saharan Africa, there is less potential to capture spillovers because of the relative uniqueness of their agroclimatic conditions and crops (box 7.4). Ways to increase investment in R&D Increasing public funding of R&D will require greater political support to agriculture, particularly to finance public goods. Forming coalitions of producers and agribusinesses around particular commodities or value chains may be the most effective way to lobby for more public funding and for producers and agribusiness to cofinance R&D. In addition, institutional reforms, discussed next, will be needed to make investing in public R&D organizations more attractive and more effective. Another way to increase investment is to remove barriers to private investment BOX 7.4 Sub-Saharan Africa s agricultural R&D challenge In addition to stagnant R&D spending, Sub- Saharan Africa faces specific challenges that add urgency to increasing the spending on agricultural R&D, extension, and associated services: The potential to capture spillovers of technology from outside the region is less in Sub-Saharan Africa than in other regions. This is partly because the crops grown in Sub- Saharan Africa are more diverse, with many so-called orphan crops where there is little global public or private R&D (for example, cassava, yams, millet, plantain, teff ), and partly because of agroecological distance. Using an index of agroecological distance zero to represent no potential for spillovers from high-income countries, where most R&D is conducted, and 1 for perfect spillover potential Pardey and others (2007) estimate that the average index for African countries is 0.05, compared with 0.27 for all developing countries. So, technologies imported from other continents often do not perform well. There is considerable heterogeneity within Africa resulting from rainfed production systems, reducing the spillover potential among countries in the region. Because of small country size, agricultural research systems in Sub-Saharan Africa are fragmented into nearly 400 distinct research agencies, nearly four times the number in India and eight times that in the United States (table below). This prevents realizing economies of scale in research. Funding per scientist is especially low in Sub-Saharan Africa. With nearly 50 percent more scientists than India, and about a third more than the United States, all of Sub-Saharan Africa spends only about half of what India spends and less than a quarter of what the United States spends. Only a quarter of African scientists have a PhD, compared with all or most scientists in India and the United States. Complex agricultural challenges in Sub- Saharan Africa require combining genetic improvement emphasizing pests, diseases, and drought, with improvements in soil and water management, and with labor-saving technologies in areas of low population density or serious HIV/AIDS infection. These problems are surmountable. First, Australia, another dryland continent technologically distant from other regions, has one of the highest intensities of public R&D investment in the world (more than 4 percent of agricultural GDP); it has a productive and competitive agricultural sector. Second, spillovers can be better targeted at a world scale for example, East African highland countries such as Ethiopia and Kenya have product mixes and agroecological conditions similar to Mexico. Third, the rise of regional research organizations in Africa should help achieve economies of scale and scope. Comparison of research systems in Sub-Saharan Africa, India, and the United States around 2000 Sub-Saharan Africa India United States Arable and permanent crop area (hectares, millions) Number of public agricultural research agencies Number of full-time equivalent scientists 12,224 8,100 9,368 Percentage of scientists with PhD Annual public spending on agricultural R&D (1999 int l $, millions) 1,085 1,860 3,465 Spending per scientist (1999 int l $, thousands) Sources: FAO 2006a. Pal and Byerlee 2006; Pardey and others 2007.

12 Innovating through science and technology 169 in R&D. One constraint to private R&D investment is a weak investment climate for private investors generally (see focus D). A second is weak demand from smallholders for improved technologies because of risks, credit constraints, and poor access to information. A third is that production systems and technologies in much of the developing world make it difficult to enforce IPRs. Added to these three are restrictions on private sector imports of technologies and high regulatory barriers to the release of new technologies, such as the varieties developed by the private sector. 63 More could be done to stimulate private investment in R&D by improving the environment for private innovation say, through stronger IPRs for inventions for commercial crops (see box 7.3) and lower barriers to the import and testing of technologies. Another approach is to make public funding for R&D contestable and open to private firms to implement the research, usually with private cofinancing. Competitive funding has become common, especially in Latin America, and some funds have the specific objective of funding private innovation (FONTEC in Chile, for example). Yet another approach is to establish a purchase fund or prize to reward developers of specific technologies, such as varieties resistant to a particular disease. 64 Prizes were used historically to promote inventions, such as an accurate way to measure longitude. 65 The reward could also be tied to the economic benefits actually generated. 66 Institutional arrangements to increase the efficiency and effectiveness of R&D systems Although public research organizations dominate in most developing countries, their efficiency and effectiveness in today s changing world are in question. Institutional reforms of public R&D were addressed in World Development Report They include creating well-governed autonomous bodies or public corporations, such as EMBRAPA (the Brazilian public agricultural research corporation); improving their effectiveness in assessing and responding to farmer demands; and increasing the contestability of funding through competitive funding mechanisms. To succeed, these reforms have to be accompanied by a long-term commitment to build capacity (box 7.5), which has paid off in the now-strong public research systems in Brazil, China, and India. A challenge for public research systems in Africa is attracting and retaining scientists, who operate in a global marketplace, especially women scientists who make up only 21 percent of the total (see focus G). 67 Research universities are also underused for publicly supported science. Competitive funding mechanisms for public funds have increased the role of universities in agricultural R&D in some countries. For example, percent of the competitive grants for agricultural R&D in Brazil, Chile, Ecuador, and Mexico have been channeled to universities. 68 Moreover, universities prepare the next generation of scientists. A comprehensive agricultural science policy is needed to address continuing weaknesses in university systems, especially in agriculture-based countries (see focus G). While investment in public R&D organizations remains important, the public sector cannot do it alone. Science-driven and linear research-extension-farmer approaches in which public research systems generate technologies disseminated BOX 7.5 Long-term capacity development in Ghana The Ghana Grains Development Project is one of the few African success stories of long-term donor support to strengthen national research and extension for food production. Ghana is also one of the few countries with sustained increases in per capita food production. The project focused primarily on increasing the output of maize and cowpeas through welladapted varieties and management practices for each of Ghana s agroecological zones. A special feature was the graduatelevel training of about 50 scientists, nearly all of whom returned to the project. Annual maize production jumped from 380,000 tons in 1979, when the project started, to more than 1 million tons by the project s end in Maize yields increased by 40 percent from 1.1 tons per hectare to 1.5 tons. The project s bottom-up approach integrated farmers in all stages of research and included socioeconomic assessment of the technology. Complemented by largescale extension programs supported by the NGO Sasakawa Global 2000, more than half of all maize farmers in Ghana adopted improved varieties, fertilizer, and planting methods by But after the removal of fertilizer subsidies, fertilizer use dropped to 25 percent, challenging the approach s sustainability. Adoption by women farmers (39 percent) was significantly lower than that for men (59 percent), reflecting differences in access to assets and services, and especially the biases in extension. Sources: Canadian International Development Agency, personal communication, 2006; Morris, Tripp, and Dankyi 1999.